User Stances Research Articles

Exploring the Opinion Shifting of Social Media Users Based on Stance Detection

ABSTRACTFine‐grained mining of stance change among social media users during crises contributes significantly to a comprehensive understanding of the development of public opinion online. This study collected detailed information on 227,281 Twitter users who continuously participated in the discussion about the COVID‐19 vaccine to analyze patterns of opinion shifting. Utilizing the COVID‐Twitter‐BERT pre‐trained model, we detect each user's stance on the COVID‐19 vaccine and construct a time‐series dataset reflecting their stance over time. A prediction model for users' opinion shifting is established based on the LightGBM model, and the SHAP explanation method is used to rank the importance of features to identify key features influencing the opinion shifting prediction. Our research findings indicate that nearly half of the users maintained a consistent stance on vaccines, with a relatively low proportion of shifts between pro‐vaccine and anti‐vaccine stances. The F1 value of the prediction model based on the LightGBM model is 0.92. These findings can assist in monitoring social media user opinions online.

Evaluating large language models for user stance detection on X (Twitter)

Current stance detection methods employ topic-aligned data, resulting in many unexplored topics due to insufficient training samples. Large Language Models (LLMs) pre-trained on a vast amount of web data offer a viable solution when training data is unavailable. This work introduces Tweets2Stance - T2S, an unsupervised stance detection framework based on zero-shot classification, i.e. leveraging an LLM pre-trained on Natural Language Inference tasks. T2S detects a five-valued user’s stance on social-political statements by analyzing their X (Twitter) timeline. The Ground Truth of a user’s stance is obtained from Voting Advice Applications (VAAs). Through comprehensive experiments, a T2S’s optimal setting was identified for each election. Linguistic limitations related to the language model are further addressed by integrating state-of-the-art LLMs like GPT-4 and Mixtral into the T2S framework. The T2S framework’s generalization potential is demonstrated by measuring its performance (F1 and MAE scores) across nine datasets. These datasets were built by collecting tweets from competing parties’ Twitter accounts in nine political elections held in different countries from 2019 to 2021. The results, in terms of F1 and MAE scores, outperformed all baselines and approached the best scores for each election. This showcases the ability of T2S, particularly when combined with state-of-the-art LLMs, to generalize across different cultural-political contexts.

Zero-shot stance detection based on multi-expert collaboration

Zero-shot stance detection is pivotal for autonomously discerning user stances on novel emerging topics. This task hinges on effective feature alignment transfer from known to unseen targets. To address this, we introduce a zero-shot stance detection framework utilizing multi-expert cooperative learning. This framework comprises two core components: a multi-expert feature extraction module and a gating mechanism for stance feature selection. Our approach involves a unique learning strategy tailored to decompose complex semantic features. This strategy harnesses the expertise of multiple specialists to unravel and learn diverse, intrinsic textual features, enhancing transferability. Furthermore, we employ a gating-based mechanism to selectively filter and fuse these intricate features, optimizing them for stance classification. Extensive experiments on standard benchmark datasets demonstrate that our model significantly surpasses existing baseline models in performance.

Incorporating target-aware knowledge into prompt-tuning for few-shot stance detection

Stance detection, a fundamental task in natural language processing, identifies user stances in texts towards specific targets. The diverse targets and ever-changing expressions make it challenging to attain comprehensive knowledge from limited data. Existing methods focus on incorporating supplementary knowledge, neglecting fusion consistency during training, which is critical for preserving the rationality of the inference. In this paper, we introduce TAP, a novel approach for few-shot stance detection. TAP extends the verbalizer hierarchically, a mapping function in prompt-tuning. Constructed using a log-odds ratio of topics and targets, the verbalizer refines candidates with prior knowledge, forming the foundation for subsequent hierarchical distillation. The hierarchical distillation, a technique based on pilot experiments on the hierarchical verbalizer, ensures the fusion of diverse knowledge during prompt-tuning, maintaining consistency throughout the training process. Notably, TAP constructs verbalizers without external knowledge augmentation. The hierarchical distillation involves a joint loss function, contributing to the model’s robustness and training consistency. Extensive experiments are conducted on SemEval2016t6 and ArgMin datasets with 13 different targets. The proposed method is evaluated on various few-shot and full-data settings with F1-Macro and F1-Micro scores. On average, TAP achieves overall improvements of 4.71% and 3.76% over state-of-the-art baselines on Semeval2016t6 and ArgMin datasets, respectively, in few-shot scenarios.

Quantifying Variations in Controversial Discussions within Kuwaiti Social Networks

During the COVID-19 pandemic, pro-vaccine and anti-vaccine groups emerged, influencing others to vaccinate or abstain and leading to polarized debates. Due to incomplete user data and the complexity of social network interactions, understanding the dynamics of these discussions is challenging. This study aims to discover and quantify the factors driving the controversy related to vaccine stances across Kuwaiti social networks. To tackle these challenges, a graph convolutional network (GCN) and feature propagation (FP) were utilized to accurately detect users’ stances despite incomplete features, achieving an accuracy of 96%. Additionally, the random walk controversy (RWC) score was employed to quantify polarization points within the social networks. Experiments were conducted using a dataset of vaccine-related retweets and discussions from X (formerly Twitter) during the Kuwait COVID-19 vaccine rollout period. The analysis revealed high polarization periods correlating with specific vaccination rates and governmental announcements. This research provides a novel approach to accurately detecting user stances in low-resource languages like the Kuwaiti dialect without the need for costly annotations, offering valuable insights to help policymakers understand public opinion and address misinformation effectively.

Neural network approaches for rumor stance detection: Simulating complex rumor propagation systems

SummaryThis research introduces a comprehensive suite of neural network models designed to tackle the challenging task of rumor stance detection within the framework of simulating complex rumor propagation systems. Our objective centers on accurately modeling the intricate structures of rumor dialogues and propagation patterns to identify user stances—whether they are in support, denial, questioning, or commenting on rumors. Unlike conventional methods that rely on simplistic keyword targeting and fail in the nuanced context of social networks, our models delve into the complexities of dialogue and propagation structures, offering a more precise and insightful analysis of rumor dynamics. In addressing the simulation and modeling of complex systems, our approach specifically focuses on the elaborate interaction networks that underpin rumor spread and reception. While our methodology does not directly engage with brain‐like computing paradigms, it reflects a similar level of sophistication in handling layered and complex information flows, analogous to cognitive processes in understanding and interpreting human communications. Employing a hierarchical attention mechanism, our models adeptly parse through multitiered dialogue sequences, effectively distinguishing between various indicators of user stances. This allows for a nuanced and detailed representation of the rumor ecosystem, significantly enhancing the accuracy of stance detection. Through rigorous testing on diverse datasets, our approach has demonstrated superior performance over existing models, thereby establishing a new benchmark in the field.

Infusing external knowledge into user stance detection in social platforms

Stance detection for user reviews on social platforms aims to classify the stance of users’ reviews toward a specific topic. Existing studies focused on the internal semantic features of reviews’ texts, but ignored the external knowledge associated with the review. This paper retrieves external knowledge related to the key information of each review by mapping it to a knowledge graph. Thereafter, this paper infuses the external knowledge into deep learning model for stance detection. Considering that infusing external knowledge may bring noise to the model, this paper adopts the personalized PageRank method to filter the introduced irrelevant external knowledge. Infusing external knowledge can improve the classification performance by providing background knowledge. In addition to considering the textual features of reviews when constructing the stance detection model, this paper employs a gated graph neural network (GGNN) approach to fuse the structural information between reviews to capture the interactions of reviews. The experiments show that the model improves 1.5% –6.9% in macro-average scores compared to six benchmark models in this paper. By combining the textual features and structural information of reviews and introducing external knowledge, the model effectively improves the stance detection performance.

Enhancing prediction of user stance for social networks rumors

The spread of social media has led to a massive change in the way information is dispersed. It provides organizations and individuals wider opportunities of collaboration. But it also causes an emergence of malicious users and attention seekers to spread rumors and fake news. Understanding user stances in rumor posts is very important to identify the veracity of the underlying content as news becomes viral in a few seconds which can lead to mass panic and confusion. In this paper, different machine learning techniques were utilized to enhance the user stance prediction through a conversation thread towards a given rumor on Twitter platform. We utilized both conversation thread features as well as features related to users who participated in this conversation, in order to predict the users’ stances, in terms of supporting, denying, querying, or commenting (SDQC), towards the source tweet. Furthermore, different datasets for the stance-prediction task were explored to handle the data imbalance problem and data augmentation for minority classes was applied to enhance the results. The proposed framework outperforms the state-of-the-art results with macro F1-score of 0.7233.

An influences-adapted two-phase approach to stance detection in the diachronic perspective

Stance detection in social networks is still a challenging research topic, and in recent years, various methods and techniques have been proposed to deal with it. The main challenge in stance detection is that the content does not directly refer to the target. Most of the conducted research have used side information such as user interactions, target phrase, and external knowledge to increase the efficiency of their method. The stance of the users changes and evolves over time, so stance detection is proposed from the diachronic perspective. Two crucial challenges of detection in the diachronic perspective are to derive the evolution and adaptation of the detection system. In this article, we deal with stance detection from the diachronic perspective by extracting the evolution of the users' stances with the help of side information and adapting with evolution. Our approach is a two-phase method. In the first phase, by using subjectivity detection, the times when each user has taken a passive stance towards the target are determined. In the second phase, by using an influence model, the evolution of the active stance of the user is extracted, and its adaptation is made. The proposed method uses tweets from user’s home timeline and the user's favorite list as side information to discover the evolution. It can adapt itself to the dynamics of users' stances evolution by using the evolution of users' stances and automatically determine better time intervals for dividing tweets and improve user stance tracking. For the first time, we performed stance detection in the diachronic perspective on the SemEval-2016 Task 6A benchmark dataset and achieved Favg=78.18. Two other datasets are also used for experiments, and the results show that adaption increases the F-score between 3.99 and 9.72%.

A multilayered graph-based framework to explore behavioural phenomena in social media conversations

ObjectiveSocial media is part of current health communications. This research aims to delve into the effects of social contagion, biased assimilation, and homophily in building and changing health opinions on social media. Materials and methodsConversations about COVID-19 vaccination on English and Spanish Twitter are the case studies. A new multilayered graph-based framework supports the integrated analysis of content similarity within and across posts, users, and conversations to interpret contrasting and confluent user stances. Deep learning models are applied to infer stance. Graph centrality and homophily scores support the interpretation of information reproduction. ResultsThe results show that semantically related English posts tend to present a similar stance about COVID-19 vaccination (rstance = 0.51) whereas Spanish posts are more heterophilic (rstance = 0.38). Neither case showed evidence of homophily regarding user influence or vaccine hashtags. Graph filters for Pfizer and Astrazeneca with a similarity threshold of 0.85 show stance homophily in English scenarios (i.e. rstance = 0.45 and rstance = 0.58, respectively) and small homophily in Spanish scenarios (i.e. r = 0.12 and r = 0.3, respectively). Highly connected users are a minority and are not socially influential. Spanish conversations showed stance homophily, i.e. most of the connected conversations promote vaccination (rstance = 0.42), whereas English conversations are more likely to offer contrasting stances. ConclusionThe methodology proposed for quantifying the impact of natural and intentional social behaviours in health information reproduction can be applied to any of the main social platforms and any given topic of conversation. Its effectiveness was demonstrated by two case studies describing English and Spanish demographic and sociocultural scenarios.

An Improved BiLSTM Approach for User Stance Detection Based on External Commonsense Knowledge and Environment Information

In the age of social networks, the number of tweets sent by users has led to a sharp rise in public opinion. Public opinions are closely related to user stances. User stance detection has become an important task in the field of public opinion. However, previous studies have not distinguished between user viewpoints and stances. These studies usually detected stance from the perspective of the tweet level but rarely the user level. Therefore, in this paper, we defined user stance, which is the user viewpoint (support, oppose, and neutral) toward the entire target event process. On this basis, we put forward a user stance detection method based on external commonsense knowledge (such as SenticNet) and environment information (such as a user’s historical tweets, topic information, and neighbor tweets) and denote this method as ECKEI. First, in order to better integrate external commonsense knowledge into the neural network, we improved BiLSTM and called it CK-BiLSTM for complementary commonsense information to the memory cell. Secondly, we used LDA to extract the topic of user tweets and designed a topic-driven module to capture the users’ neighbors’ information. Finally, we used the attention mechanism to integrate information from users’ historical tweets and neighbors’ tweets obtained through topic information; then, we used the softmax layer to classify user stances into the support, neutral and oppose classes. In this paper, we conducted experiments and assessments on datasets containing information on Brexit and the elections to verify the practicability and effectiveness of our proposed method. Extensive experimental results on the Brexit and elections datasets show that our approach outperforms six baseline methods (SVM-ngram, NB, MTTRE (RNN), Pkudblab (CNN), TAAT, and Aff-feature). We use the average micro-F1 and average accuracy to measure performance on the detection of a user’s stance. The ECKEI model makes improvements of 4.30–16.89% and 1.22–16.58% on the Brexit and election datasets, respectively, in terms of average micro-F1. Meanwhile, this model makes improvements of 4.24–17.46% and 0.48–14.64% on the Brexit and election datasets, respectively, in terms of average accuracy. Our model makes improvements of 5.34–17.30% and 2.65–19.73%, respectively, on the Brexit and election datasets in terms of average recall.

Rumor Stance Classification in Online Social Networks: The State-of-the-Art, Prospects, and Future Challenges

The emergence of the Internet as a ubiquitous technology has facilitated the rapid evolution of social media as the leading virtual platform for communication, content sharing, and information dissemination. In spite of revolutionizing the way news is delivered to people, this technology has also brought along with itself inevitable demerits. One such drawback is the spread of rumors expedited by social media platforms, which may provoke doubt and fear. Therefore, it is essential to debunk rumors before their widespread use. Over the years, many studies have been conducted to develop effective rumor verification systems. One aspect of such studies focuses on rumor stance classification, which involves the task of utilizing user viewpoints regarding a rumorous post to better predict the veracity of a rumor. Relying on user stances in rumor verification has gained significant importance, for it has resulted in significant improvements in the model performance. In this paper, we conduct a comprehensive literature review of rumor stance classification in complex online social networks (OSNs). In particular, we present a thorough description of these approaches and compare their performances. Moreover, we introduce multiple datasets available for this purpose and highlight their limitations. Finally, challenges and future directions are discussed to stimulate further relevant research efforts.

Evaluating Rumor Debunking Effectiveness During the COVID-19 Pandemic Crisis: Utilizing User Stance in Comments on Sina Weibo.

Background: The spread of rumors related to COVID-19 on social media has posed substantial challenges to public health governance, and thus exposing rumors and curbing their spread quickly and effectively has become an urgent task. This study aimed to assist in formulating effective strategies to debunk rumors and curb their spread on social media.Methods: A total of 2,053 original postings and 100,348 comments that replied to the postings of five false rumors related to COVID-19 (dated from January 20, 2020, to June 28, 2020) belonging to three categories, authoritative, social, and political, on Sina Weibo in China were randomly selected. To study the effectiveness of different debunking methods, a new annotation scheme was proposed that divides debunking methods into six categories: denial, further fact-checking, refutation, person response, organization response, and combination methods. Text classifiers using deep learning methods were built to automatically identify four user stances in comments that replied to debunking postings: supporting, denying, querying, and commenting stances. Then, based on stance responses, a debunking effectiveness index (DEI) was developed to measure the effectiveness of different debunking methods.Results: The refutation method with cited evidence has the best debunking effect, whether used alone or in combination with other debunking methods. For the social category of Car rumor and political category of Russia rumor, using the refutation method alone can achieve the optimal debunking effect. For authoritative rumors, a combination method has the optimal debunking effect, but the most effective combination method requires avoiding the use of a combination of a debunking method where the person or organization defamed by the authoritative rumor responds personally and the refutation method.Conclusion: The findings provide relevant insights into ways to debunk rumors effectively, support crisis management of false information, and take necessary actions in response to rumors amid public health emergencies.

Embeddings-Based Clustering for Target Specific Stances: The Case of a Polarized Turkey

On June 24, 2018, Turkey conducted a highly consequential election in which the Turkish people elected their president and parliament in the first election under a new presidential system. During the election period, the Turkish people extensively shared their political opinions on Twitter. One aspect of polarization among the electorate was support for or opposition to the reelection of Recep Tayyip Erdoğan. In this paper, we present an unsupervised method for target-specific stance detection in a polarized setting, specifically Turkish politics, achieving 90% precision in identifying user stances, while maintaining more than 80% recall. The method involves representing users in an embedding space using Google's Convolutional Neural Network (CNN) based multilingual universal sentence encoder. The representations are then projected onto a lower dimensional space in a manner that reflects similarities and are consequently clustered. We show the effectiveness of our method in properly clustering users of divergent groups across multiple targets that include political figures, different groups, and parties. We perform our analysis on a large dataset of 108M Turkish election-related tweets along with the timeline tweets of 168k Turkish users, who authored 213M tweets. Given the resultant user stances, we are able to observe correlations between topics and compute topic polarization.

Predicting Stance Polarity and Intensity in Cyber Argumentation With Deep Bidirectional Transformers

In online deliberation, participants argue in support or opposition to one another's arguments and ideas to advocate their position. Often their stance expressed in their posts is implicit and must be derived from the post's text. Existing stance detection models predict the polarity of the user's stance from the text, but do not consider the stance's intensity. We introduce a new research problem, stance polarity, and intensity prediction in response relationships between posts. This problem seeks to predict both the stance polarity and intensity of a replying post toward its parent post in online deliberation. Using our cyber argumentation platform, we have collected an empirical data set with explicitly labeled stance polarity and intensity relationships. In this work, we create six models: five are adapted from top-performing stance detection models and another novel model that fine-tunes the deep bidirectional transformer model BERT. We train and test these six models on our empirical data set to compare their performance for stance polarity and intensity prediction and stance detection. Our results demonstrate that our method of encoding the stance polarity and intensity labels allows the models to predict stance polarity and intensity without compromising their accuracy for stance detection, making these models more versatile. Our results reveal that a novel split architecture for fine-tuning the BERT model outperforms the other models for stance polarity and intensity prediction by 5% accuracy. This work is the first to train models for predicting both the stance polarity and intensity in one combined task while maintaining good accuracy.

Unsupervised User Stance Detection on Twitter

We present a highly effective unsupervised framework for detecting the stance of prolific Twitter users with respect to controversial topics. In particular, we use dimensionality reduction to project users onto a low-dimensional space, followed by clustering, which allows us to find core users that are representative of the different stances. Our framework has three major advantages over pre-existing methods, which are based on supervised or semi-supervised classification. First, we do not require any prior labeling of users: instead, we create clusters, which are much easier to label manually afterwards, e.g., in a matter of seconds or minutes instead of hours. Second, there is no need for domain- or topic-level knowledge either to specify the relevant stances (labels) or to conduct the actual labeling. Third, our framework is robust in the face of data skewness, e.g., when some users or some stances have greater representation in the data. We experiment with different combinations of user similarity features, dataset sizes, dimensionality reduction methods, and clustering algorithms to ascertain the most effective and most computationally efficient combinations across three different datasets (in English and Turkish). We further verified our results on additional tweet sets covering six different controversial topics. Our best combination in terms of effectiveness and efficiency uses retweeted accounts as features, UMAP for dimensionality reduction, and Mean Shift for clustering, and yields a small number of high-quality user clusters, typically just 2–3, with more than 98% purity. The resulting user clusters can be used to train downstream classifiers. Moreover, our framework is robust to variations in the hyper-parameter values and also with respect to random initialization.

Your Stance is Exposed! Analysing Possible Factors for Stance Detection on Social Media

To what extent user's stance towards a given topic could be inferred? Most of the studies on stance detection have focused on analysing user's posts on a given topic to predict the stance. However, the stance in social media can be inferred from a mixture of signals that might reflect user's beliefs including posts and online interactions. This paper examines various online features of users to detect their stance towards different topics. We compare multiple set of features, including on-topic content, network interactions, user's preferences, and online network connections. Our objective is to understand the online signals that can reveal the users' stance. Experimentation is applied on tweets dataset from the SemEval stance detection task, which covers five topics. Results show that stance of a user can be detected with multiple signals of user's online activity, including their posts on the topic, the network they interact with or follow, the websites they visit, and the content they like. The performance of the stance modelling using different network features are comparable with the state-of-the-art reported model that used textual content only. In addition, combining network and content features leads to the highest reported performance to date on the SemEval dataset with F-measure of 72.49%. We further present an extensive analysis to show how these different set of features can reveal stance. Our findings have distinct privacy implications, where they highlight that stance is strongly embedded in user's online social network that, in principle, individuals can be profiled from their interactions and connections even when they do not post about the topic.

We Like, We Post: A Joint User-Post Approach for Facebook Post Stance Labeling

Web post and user stance labeling is challenging not only because of the informality and variation in language on the Web but also because of the lack of labeled data on fast-emerging new topics—even the labeled data we do have are usually heavily skewed. In this paper, we propose a joint user-post approach for stance labeling to mitigate the latter two difficulties. In labeling post stance, the proposed approach considers post content as well as posting and liking behavior, which involves users. Sentiment analysis is applied to posts to acquire their initial stance, and then the post and user stance are updated iteratively with correlated posting-related actions. The whole process works with limited labeled data, which solves the first problem. We use real interaction between authors and readers for stance labeling. Experimental results show that the proposed approach not only substantially improves content-based post stance labeling, but also yields better performance for the minor stance class, which solves the second problem.

Animating the ethical demand

This paper addresses the challenge of attaining ethical user stances during the design process of products and services and proposes animation-based sketching as a design method, which supports elaborating and examining different ethical stances towards the user. The discussion is qualified by an empirical study of Responsible Research and Innovation (RRI) in a Triple Helix constellation. Using a three-week long innovation workshop, UCrAc, involving 16 Danish companies and organisations and 142 students as empirical data, we discuss how animation-based sketching can explore not yet existing user dispositions, as well as create an incentive for ethical conduct in development and innovation processes. The ethical fulcrum evolves around Løgstrup's Ethical Demand and his notion of spontaneous life manifestations. From this, three ethical stances are developed; apathy, sympathy and empathy. By exploring both apathetic and sympathetic views, the ethical reflections are more nuanced as a result of actually seeing the user experience simulated through different user dispositions. Exploring the three ethical stances by visualising real use cases with the technologies simulated as already being implemented makes the life manifestations of the users in context visible. We present and discuss how animation-based sketching can support the elaboration and examination of different ethical stances towards the user in the product and service development process. Finally we present a framework for creating narrative representations of emerging technology use cases, which invite to reflection upon the ethics of the user experience.

Facebook as a space for agency: negotiations by users and non-users

The Web’s architecture sets conditions for how people communicate, interact and act online in general and on social network sites (SNSs) in particular. Users and non-users of Facebook were invited to focus groups in order to create an understanding of Facebook as a technology-afforded space for agency and to illuminate the variety of meanings given to it. The aim of this study is to identify user and non-user stances to the possibilities of action and actor roles – understood in terms of mediated agency – enabled by SNSs, particularly Facebook. The negotiation of Facebook’s nature as a space is scrutinised using concepts of connectedness and connectivity as analytical lenses.